I 
The remaining outcome parameters of this study are more descriptive in 
nature. It is not possible to describe fully an analytic plan for these 
outcomes. The approach to the analysis must be interactive and will depend 
upon preliminary and descriptive results. Nevertheless, the following section 
provides a general introduction to the approach that will be used. Study 
parameters are listed in section 7.3 of the protocol. Variables measured on a 
continuous scale, such as laboratory measurements, will be characterized by 
summary statistics (mean and standard deviation) . The randomized study design 
permits a valid comparision of means between the treatment groups. However, 
because of the small sample size, we will not emphasize such comparisions . If 
they are made, standard tests such as t-tests or their non-parametric analogs 
will be performed. 
Other variables are dichotomous or categorical in outcome such as the 
results of symptom inventory, presence or absence of auto antibodies, and 
response to allogenic stimulator cells. These variables will be summarized 
using counts and proportions with exact binomial confidence limits. Here 
again, any comparisons between the treatment groups will be made using either 
Fisher's exact test or its multi-dimensional analog. 
Pharmacokinetic parameters will be estimated, when possible, using 
standard compartmental models. Parameter estimates and normal theory standard 
errors will be calculated by fitting using weighted least squares. The 
relationship between pharmacokinetic parameters and clinical outcomes such as 
toxicity will be assessed using linear regression or a related technique. 
We do not anticipate the need for detailed multivariate analysis. Also, the 
sample size employed will not support the precise estimate of covariate 
effects adjusted for other prognostic factors. 
Biostatistics References 
1. Storer, B.E. Design and Analysis of Phase I Clinical Trials. Biometrics, 
45:925-937, 1989. 
2. O’Quigley, J. Estimating the Probability of Toxicity at the Recommended 
Dose Following a Phase I Clinical Trial in Cancer. Biometrics, 48:853- 
862, 1992. 
3. O'Quigley, kJ., and Chevret, S. Methods for Dose Finding Studies in 
Cancer Clinical Trials: A Review in Results of a Monte Carlo Study. 
Statistics in Medicine, 10: 1647-1664, 1991. 
4. O'Quigley, J. , Pepe, M. and Fisher, L. Continual Reassessment Method: A 
Practical Design for Phase I Clinical Trials in Cancer. Biometrics, 
46:33-48, 1990. 
5. Gatsonis, C. and Greenhouse, J.B. Bayesian Methods for Phase I Clinical 
Trials. Statistics in Medicine, 111377-1389, 1992. 
[324] 
Recombinant DNA Research, Volume 17 
